In some image-based positioning systems, the system may match details of a single building including text on doors or style of entrance. These building details may be limiting for matching purposes since there may be a common construction theme or building design used in multiple areas of a city or country. These building details may also be difficult to detect and may have added error due to camera angle or resolution. These location systems may return too many results due to the thin matching parameters, or return few or no results due to camera angle or resolution. Additionally, detailed building features may be difficult to determine and match due to camera position, e.g. angle or focal lengths differing from a reference building image.
Some city-scale landmark identifiers determine a match probability of an image to each of the reference landmark images. It is a tremendous computational and storage burden to store and match each image to all reference images. Additionally, the match determination may be limited by necessitating the landmark to be predominate in the image. Further, changes in regions proximate to the landmark may hinder search results as the processed image may not match the background features of the reference image.